1. What can be expected from risk scores for predicting postoperative nausea and vomiting? 
C.C. Apfel, P. Kranke, C.-A. Greim, N. Roewer 
British Journal of Anaesthesia 
Volume 86, Issue 6, Pages (June 2001)
DOI: /bja/
Copyright © 2001 British Journal of Anaesthesia Terms and Conditions

2. Fig 1
ROC-curve for the single predictor models and for the four-predictor model. The single predictor model considered either female gender, non-smoking, history of motion sickness or PONV and post-operative opioids while the four-predictor model considered all predictors simultaneously. Note, that the four-predictor model had a clinically relevant larger AUC than any single predictor model.
British Journal of Anaesthesia  , DOI: ( /bja/ )
Copyright © 2001 British Journal of Anaesthesia Terms and Conditions

3. Fig 2
(Upper curve) AUC for population I, for example, when all parameters were taken from a previous study.11 The interval indicated by the vertical lines displays the range of the published AUC when four or five predictors were applied to its original data set. (Lower curve) AUC for population II, for example, when this score is applied to a gynaecological setting. Note, that considering the additional predictor such as ‘laparoscopy’ does not lead to a clinically relevant improvement. The dotted lines indicate that the increase of the AUC per additional predictor is less than MS-PONV hist=history of motion sickness or PONV. ROC-curve=receiver operating characteristic-curve.
British Journal of Anaesthesia  , DOI: ( /bja/ )
Copyright © 2001 British Journal of Anaesthesia Terms and Conditions

4. Fig 3
(Lower curve) AUC for population III, for example, all predictors were assumed to have an OR of 2. (Upper curve) AUC for population IV, for example, all predictors were assumed to have an odds ratio of 3. Note, that a four-predictor model is much better than a single factor prediction. However, from four or five factors onwards, the clinical relevance becomes questionable as the increase of the area under the curve per additional predictor becomes less than 0.025, as indicated by the dotted lines. ROC-curve=receiver operating characteristic-curve.
British Journal of Anaesthesia  , DOI: ( /bja/ )
Copyright © 2001 British Journal of Anaesthesia Terms and Conditions